Clever Geek Handbook

Are C ++ 2 bit bitmaps possible?

I have a structure of two bit bit fields:

struct MyStruct {
    unsigned __int32 info0  : 2;
    unsigned __int32 info1  : 2;
    unsigned __int32 info2  : 2;
   ...
    unsigned __int32 info59 : 2;
};

And also, how it goes up to 120 ... Is there a way to write and address them as an array?

+3


source to share


2 answers


If you cannot use Paul R's answer for any reason, you can always use a custom accessory with a standard array:

static unsigned __int8 infos[30]; // 240 bits allocated

unsigned __int8 getInfo( unsigned short id_num )
{
    return (infos[id_num/4] >> ((2*id_num) % 8) ) & 0x3;
}
// setInfo left as an exercise.


(You may need to test the logic here, I haven't tested it.)

+3


source


I would use a proxy to create a temporary reference that could be used to manipulate 2-bit elements using array syntax. This can be easily changed to handle n-bit items.



#include <iostream>

class TwoBitArray {
public:
    typedef unsigned char byte;

    TwoBitArray(unsigned size) : bits(new byte[(size + 3) / 4]) {}
    ~TwoBitArray() { delete bits; }

    class tbproxy {
    public:
        tbproxy(byte& b, int pos) : b(b), pos(pos) {}

        // getter
        operator int() const {
            return (b >> (pos * 2)) & 3;
        }

        // setter
        tbproxy operator=(int value) {
            const byte mask = ~(3 << (pos * 2));
            b = (b & mask) | (value << (pos * 2));
            return *this;
        }

    private:
        byte& b;
        int pos;
    };

    // create proxy to manipulate object at index
    tbproxy operator[](int index) const {
        return tbproxy(bits[index/4], index & 3);
    }

private:
    byte* bits;
};

int main() {
    const int size = 20;
    TwoBitArray a(size);
    for (int i = 0; i < size; ++i)
        a[i] = i & 3;
    for (int i = 0; i < size; ++i)
        std::cout << i << ": " << a[i] << std::endl;
}
+2


source






More articles:


All Articles